shrug gotta start somewhere. I’m not under any delusions we will have full self driving in the next decade but getting this use case solidified and in real commercial use (day to day, not one off) is important imo.
Why? Not saying it isn’t, but I don’t really see immediately why it is.
As a software developer my main mantra is to start with an MVP and iterate to work out from there (where feasible). So proving this can work with any unpredictable circumstances that arise (erratic traffic patterns out of nowhere, accidents, etc…) lead the way for these specific circumstances to get regulatory approval for part time drivers (instead of requiring these routes being manned every minute of operation). That then leads to economies of scale because you get approval for convoys.
The bigger thing is you start getting real FSD use cases being accepted by regulators and start getting a framework for how they evaluate FSD applications. Tesla’s plan of “I’ll have FSD ready by the end of the year” is laughable because states don’t even have an idea how to proceed for determining if it’s ready or not and trying to have it nationally available at one time ignores the vast amount of edge cases in the entirety of the US.
Not only do these commercial use cases start the ball rolling it’s being done by companies that have real lobbying power and financial backing.
As successful day to day operations prove to be successful and they gain more telemetry and insight into how to expand operation to other highways that are still ideal but not as perfect conditions as the Phoenix ↔ Tuscany route and slowly expand from there.
So, basically they have become trains? But less efficient?
Just going along a single route?
And they have doubled the number of people in the cab versus regular trucks!
My feeling is that autonomous vehicle development in general has hit a plateau just short of Level 4, and that there is no clear path for advancement beyond that plateau.
I was really asking what you thought the end game was. I see a possible role for self-driving for depot-to-depot transport, but even that seems dicey to me. And as I alluded to, that sort of application just recreates railroads with far worse than economies of scale.
I kind of feel the same way. Best case, they can do limited-planned-route trips, but buses and trains do that far better.
Such a great movie. Who said cocaine was bad?
Depends what you mean by end game, and what you mean by worse economies of scale. Self driving trucks on highways can allow for a really efficient distribution system with a convoy of trucks going between distribution hubs while only needing 1 driver.
You say “recreates railroads with less economies of scale” but our rail road system is pretty bad in the US and is pretty limited in where it goes. You also don’t have to worry about scheduling time and space with existing rail freight trains (since you own the trucks you can send them off as needed instead of waiting for scheduled times). It’s probably much easier to send a convoy to a distribution hub and just have each truck go their own way (manually driven) once at the distribution hub than to have to unload the trains and load them right onto trucks to go out.
To me it seems like there’s a lot of potential for efficiency in that system.
It’s bad because we chose to abandon it for cars and trucks and highways. That’s a choice, not an inevitability.
I could be wrong, but I don’t think there is much controversy in the idea that rail freight costs are dramatically cheaper than truck freight costs.
I mean if you want to have a discussion that the US needs more rail support then you aren’t going to get any argument from me. However, from a practicality point of view that’s never going to happen because it’s too expensive to lay down new track at this point (not only in the actual costs to execute it but also for buying up land, dealing with NIMBY-ism, etc…). It’s not really economical currently for an individual company to do that, and there’s zero political will (or civilian will) for that to occur.
So for the time being it’s more efficient to use our current infrastructure (highways) to streamline distribution networks than to create new ones.
I don’t have skin in the game so I don’t gain much by freight becoming autonomous, but I don’t blame them for taking that route either.
Regulations limit the number of hours a driver can drive. It’s all logged. Which means that drivers often spend huge amounts of time driving as fast as legally possible to cover as much ground as possible before they have to shut it down and wait for their time limit to reset.
Another way to think of it is assume a human driver drives 16 hours per day, but needs 8 for food, bathroom, sleep, etc. At 65 mph, they’re doing 1,040 miles per day. But a driverless semi could do 45 mph for 24 hours and do 1,080 miles per day. And the kicker is that they’re burning a lot less fuel at 45 mph than at 65 mph, because going faster with that much weight always translates into burning more fuel. And since they’re also going slower, they’re also safer for everyone else on the highway.
Sort of… Not sure if building another rail line connecting these two cities is cheaper than self-driving trucks on existing roads. If these trucks don’t run over people then I would say they are a limited success.
The “standby” drivers (I forget the correct term) are kind of a sick joke though. To me it’s sort of like letting the drunk guy drive and having the designated driver in the passenger seat.
The thing is, there is already an unused rail line between Tucson and Phoenix. Probably has to be upgraded — it hasn’t been used in decades — but all the hard work of dealing with the terrain has been done. Just as it has been with the highway, which — if it did not exist — would have to be built for the trucks to use.
This is a common misconception—the US passenger rail system isn’t very good, but the US freight rail system is among the world’s best. In terms of tons moved as a percentage of total internal freight tonnage, US rail carriers notch a figure of about 40%. Europe is stuck down in the teens and twenties, even with their more extensive rail networks. (Largely because their rail systems are passenger-focused, which is sexier but represents a much smaller efficiency gain than replacing tractor-trailers.)
I have to leave in seven minutes, so I can’t find the sources I hit when I first generated this claim a month or two ago, but if you don’t believe me, tag me, and I’ll see if I can dig them up this afternoon.
I wasn’t sure about the world’s best, but my understanding is freight moving here is actually very good. Most people don’t think too much about how stuff gets tot he shelves though, and when they do they just think trucks. They forget about ships, until the news reports about a specific one sinking, and airplanes and yes, trains too.
Even though this area used to be known for passenger trains, we don’t have any here anymore, but the lines are still used by freight.
Freight transport is a much bigger part of the nation’s infrastructure than most people realize. In 2013, about $50 billion worth of goods was transported as freight cargo every single day in the United States alone. Trucks move 29 percent of the freight ton-miles, but are responsible for 77 percent of the sector’s emissions.
Using robot trucks to move freight is inefficient compared to trains, but the idea of robot cars as mass transit is even more inefficient. Advocates don’t seem to consider loading factor or traffic analysis at all.
I’m deeply skeptical that fully autonomous vehicles will even work. They may work for narrowly defined applications (geofenced rigidly controlled routes, like a bus or shuttle line, or a train between two warehouses), but they will not work technically outside of those narrow applications; and even within them, they are not close to the same level of efficiency as mass transit or freight rail. How many cars do you need to e.g. replace the NY or London or Paris transit system? The answer is, you can’t approach the load factor those systems deliver without massive amounts of cars, and there is no place to put all those cars. Try loading a thousand people into cars in 20 seconds.
Yeah Scott, I agree. Private passenger autonomous vehicles are entirely about making driving easier and perhaps more pleasant. There might labor cost saving benefits for some industries, but there is no particular reason to believe less fuel will be consumed.
Actually, common sense should dictate that more carbon is going to be released into the atmosphere due to autonomous vehicles. At no point in American history has making driving cheaper or easier resulted in less driving. Expand a crowded highway, and more individuals take that highway. The ability to now have cars on the roads without a driver? There will be more cars on the road! Furthermore, American workers will be willing to tolerate longer commutes if that commute can be spent either on laptops working, or perhaps just watching Youtube videos on smartphones.
For all the discussion on when the technology will be ready, we aren’t having conversations on whether we’re ready for the technology!
I thought the idea was that self driving cars would all be owned by taxi companies, and that private ownership would be expansive and dead.
One of the bright sides would be the removal of all those parking lots and spaces. If the car is fully autonomous, it could drop people off in front of a store, and drive off to the next passenger. In the mean time no Walmart parking lot necessary. Instead, more green spaces, more pedestrian parks. More downtowns designed for walking and biking instead of metal machines.
People are buying Teslas with to-be-delivered-in-the-future self-driving technology on the premise that their car will earn money for them while they are at work or sleeping.
Never mind that the technology itself will never work. What happens when all of those cars, instead of being parked somewhere, are on the road circulating searching for passengers?
We already know that adding Uber and Lyft — i.e. cheap car services — has led to more traffic, not less. Those services basically cannibalize mass transit. People shift from riding buses / subways to riding alone in cheap taxis. That’s…not good.